Schizophrenia (SCZ) is a severely disabling and highly heritable condition that afflicts ~1% of the global population. Despite the advances in genetics that have shed light on the biology of so many diseases and led to groundbreaking new treatments for conditions from blindness to cancer, genetic studies in SCZ have been hindered by its genetic complexity and phenotypic heterogeneity: two patients who do not share a single symptom can both receive the same diagnosis. One way forward is to define molecular subtypes, an approach that has been successful in oncology and other realms of medicine, but which often requires identifying distinct phenotypic subtypes. We therefore propose to study the most severe form of the disorder?extremely treatment-resistant SCZ (ETRS)?because rare genetic variants of large effects are likely to be enriched in this population. ETRS patients are the top 1% most severely affected patients, often remain hospitalized for decades, and are not usually included in genetic studies, in part because they are not accessible to most researchers. We will recruit 400 subjects with ETRS from the New York State inpatient system, thoroughly characterize their phenotype, and perform whole genome sequencing (WGS). Our preliminary data from 75 ETRS patients revealed a higher-than-expected prevalence of seizures, dysmorphic features, and cognitive impairment, which suggest the presence of one of the 60 Mendelian diseases known to mimic SCZ. Furthermore, prior research and our preliminary data indicate that SCZ severity, cognitive impairment, and treatment resistance are associated with a greater burden of rare single nucleotide variants (SNVs), copy number variants (CNVs), and common variant polygenic risk. Therefore, in Aim 1, we will use diagnostic WGS to identify (a) Mendelian conditions that mimic SCZ, such as Niemann-Pick disease type C, and (b) pharmacogenetic variants that reduce the efficacy of antipsychotic treatments. Patients with either type of mutation may be treatable. In Aim 2, we will (a) evaluate the burden of rare SNVs and rare CNVs. Since extreme phenotypes can also be due to an excess of common variant risk, we will (b) determine the common variant burden by calculating SCZ polygenic risk scores. We will use 6,500 individuals with typical SCZ and 165,000 healthy individuals as controls for Aims 1 and 2. The project will be conducted at Columbia University Medical Center by Anthony Zoghbi, MD, a psychiatrist with clinical expertise and a career goal of becoming an independent investigator in psychiatric genetics, focusing on SCZ. Dr. Zoghbi's comprehensive five-year training plan will enable him to develop expertise in statistical genetics and genomics, cognitive assessment of SCZ, and clinical research methods under the mentorship of David Goldstein, PhD (Director, Institute for Genomic Medicine), statistical geneticist Suzanne Leal, PhD, and neuropsychologist Terry Goldberg, PhD. The proposed aims align with his training goals and will elucidate the genetic architecture of ETRS, shed light on the pathophysiology of SCZ, and improve the care of those most disabled by SCZ.